This paper describes a new method for recognizing hand configurations of the Brazilian Gesture Language - LIBRAS - using depth maps obtained with a Kinect® camera. The proposed method comprised three phases: hand segmentation, feature extraction, and classification. The segmentation phase is independent from the background and depends only on pixel depth information. Using geometric operations and numerical normalization, the feature extraction process was done independent from rotation and translation. The features are extracted employing two techniques: (2D)2LDA and (2D)2PCA. The classification is made with a novelty classifier. A robust database was constructed for classifier evaluation, with 12,200 images of LIBRAS and 200 gestures of each hand configuration. The best accuracy obtained was 95.41%, which was greater than previous values obtained in the literature.
Description
A new method for recognizing hand configurations of Brazilian gesture language - IEEE Conference Publication
%0 Conference Paper
%1 filho2016method
%A Filho, Cicero Ferreira Fernandes Costa
%A dos Santos, Bárbara Lobato
%A de Souza, Robson Silva
%A dos Santos, Jonilson Roque
%A Costa, Marly Guimarães Fernandes
%B 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)
%D 2016
%I IEEE
%K assistive-technology feature-extraction gesture-recognition image-segmentation mathematical-model real
%P 3829-3834
%R 10.1109/EMBC.2016.7591563
%T A new method for recognizing hand configurations of Brazilian gesture language
%U https://ieeexplore.ieee.org/abstract/document/7591563
%X This paper describes a new method for recognizing hand configurations of the Brazilian Gesture Language - LIBRAS - using depth maps obtained with a Kinect® camera. The proposed method comprised three phases: hand segmentation, feature extraction, and classification. The segmentation phase is independent from the background and depends only on pixel depth information. Using geometric operations and numerical normalization, the feature extraction process was done independent from rotation and translation. The features are extracted employing two techniques: (2D)2LDA and (2D)2PCA. The classification is made with a novelty classifier. A robust database was constructed for classifier evaluation, with 12,200 images of LIBRAS and 200 gestures of each hand configuration. The best accuracy obtained was 95.41%, which was greater than previous values obtained in the literature.
@inproceedings{filho2016method,
abstract = {This paper describes a new method for recognizing hand configurations of the Brazilian Gesture Language - LIBRAS - using depth maps obtained with a Kinect® camera. The proposed method comprised three phases: hand segmentation, feature extraction, and classification. The segmentation phase is independent from the background and depends only on pixel depth information. Using geometric operations and numerical normalization, the feature extraction process was done independent from rotation and translation. The features are extracted employing two techniques: (2D)2LDA and (2D)2PCA. The classification is made with a novelty classifier. A robust database was constructed for classifier evaluation, with 12,200 images of LIBRAS and 200 gestures of each hand configuration. The best accuracy obtained was 95.41%, which was greater than previous values obtained in the literature.},
added-at = {2019-09-11T06:45:43.000+0200},
author = {Filho, Cicero Ferreira Fernandes Costa and dos Santos, Bárbara Lobato and de Souza, Robson Silva and dos Santos, Jonilson Roque and Costa, Marly Guimarães Fernandes},
biburl = {https://www.bibsonomy.org/bibtex/23851cf06d0e12618c1d7e3a2ea6a3cf2/jpmor},
booktitle = {2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)},
description = {A new method for recognizing hand configurations of Brazilian gesture language - IEEE Conference Publication},
doi = {10.1109/EMBC.2016.7591563},
interhash = {e73c53fa8e30611daa235f4a32f7cd64},
intrahash = {3851cf06d0e12618c1d7e3a2ea6a3cf2},
issn = {978-1-4577-0220-4},
keywords = {assistive-technology feature-extraction gesture-recognition image-segmentation mathematical-model real},
language = {English},
month = {08},
pages = {3829-3834},
publisher = {IEEE},
timestamp = {2020-10-07T13:36:50.000+0200},
title = {A new method for recognizing hand configurations of Brazilian gesture language},
url = {https://ieeexplore.ieee.org/abstract/document/7591563},
year = 2016
}